Generally, a process of encoding moving pictures includes processes of image preprocessing, motion estimation and compensation, transform, quantization, variable length coding, determination of an encoding mode, a transmission rate control, and the like.
Here, in the determination of the encoding mode, a motion is estimated in an integer pixel unit and then it is determined whether a macro block is encoded with the intra mode or with the inter mode.
In the intra mode, the macro block is encoded by the processes of transform, quantization, and variable length coding. In this case, the motion estimation and compensation are not required.
On the other hand, in the inter mode, the motion estimation is performed on a current block and a motion vector (MV) is obtained as a displacement from a predicted block for the current block.
Next, after the motion is compensated from the motion vector (MV), a pixel value of the predicted block is subtracted from the current block to make a residual block and then the residual block is encoded by the same method as the intra macro block. In this case, the motion vector (MV) and texture information of the residual block are transmitted to a decoder, in which the decoder may add the quantized residual block to the predicted block by the motion vector to obtain a reconstructed image for the current block.
However, the moving pictures having less motion have a lot of temporal correlation and therefore a residual signal of the residual block is very small in a current frame.
Therefore, the number of bits required to encode the residual signal and the number of bits required to encode the motion vector are smaller than the number of bits required to encode the current block. In this case, the inter mode is appropriate.
On the other hand, when region generations, which overlap or are exposed by a rapid motion or a motion of an object, and a scene change occur, it is difficult to predict the macro block.
Therefore, since the number of bits required to encode the motion vector and the residual signal is larger than the number of bits required to encode a block of an original image, encoding the original image with the intra mode is preferable in terms of encoding efficiency and subjective image quality.
The method for determining an encoding mode according to the related art needs to include a process of comparing energy of the predicted residual signal with energy of the original image to determine the intra mode or the inter mode and therefore it takes a long encoding time due to a complicated computing process.